We propose a systematic optimization of antitumor anthracycline structures through semisynthetic derivatization of daunorubicin and doxorubicin. Specifically, we plan to pursue the important new leads provided by our recently discovered 3'-(4-morpholinyl) and 3'-(4-methoxy-1-piperidinyl) derivatives of 3'-deaminodaunorubicin. The former compound is the most potent anthracycline analog synthesized so far and the latter is an analog with markedly improved antitumor efficacy. Hence, attachment to the sugar of a new ring that incorporates the amino N and has an ether O at the 4-position appears to provide biologically important additions to the anthracycline structure that were not previously identified. Most of the further analogs will be synthesized by the Borch reductive alkylation, using various dialdehydes and sodium cyanoborohydride. Other structure changes will be based on previous leads indicating that favorably altered biological properties are obtained with the 5-imino (antitumor activity retained, less cardiotoxic) and N,N-dibenzyl derivatives (better antitumor efficacy, less cardiotoxic). A limited effort will be devoted to a totally synthetic morpholinyl analog with the aglycone simplified by deletion of the 4-methoxyl and 9-acetyl substituents. The new products will be systematically screened for antitumor properties in mice. In addition, they will be tested in vitro for DNA interactive properties and for radical-forming properties, which are relevant to the major hypothetical mechanisms of action. A limited number will be tested for cardiotoxicity in rats. Our objective is to develop better anticancer drugs by an integrated process of design, synthesis, and evaluatin, already set in operation under the non-renewable Cancer Research Emphasis grant CA 25711.